Tech Won't Save Us - The Corruption of Open Source w/ tante

Episode Date: November 7, 2024

Paris Marx is joined by tante to discuss troubling developments in the open source world as Wordpress goes to war with WP Engine and a new definition of open source AI doesn’t require being open about training data.tante is a sociotechnologist, writer, speaker, and Luddite working on tech and its social impact.Tech Won’t Save Us offers a critical perspective on tech, its worldview, and wider society with the goal of inspiring people to demand better tech and a better world. Support the show on Patreon.The podcast is made in partnership with The Nation. Production is by Eric Wickham. Transcripts are by Brigitte Pawliw-Fry.Also mentioned in this episode:tante wrote about the problem with the Open Source Initiative’s definition of open source AI.Check out this link for the full breakdown on the Wordpress drama.Wordpress changed its trademark guidelines on September 19 regarding the use of the WP abbreviation.Tumblr and Wordpress started selling user data for AI training earlier this year.A lot of the controversy around Richard Stallman started blowing up in 2019.Support the show

We all live in these digital spaces. And if it's a digital space, we should have rights, whether we can code or not. And I think that that is the shift that we need to kind of establish and that software developers kind of didn't work towards, made in partnership with The Nation magazine. I'm your host, Paris Marks, and this week my guest is Tante. You might remember I talked to Tante earlier this year. We did a live recording at the Republika conference in Berlin, but he's back on the show to talk about a pretty cool topic, I think. But before we talk about that, Tante is a sociotechnologist, writer, speaker, and Luddite working on tech and its social impact. So we've been talking a lot lately about the

politics of technology, right? In particular, looking at what these tech billionaires are up to because this has been in the discussion so much because of the US election. And while this episode is political, you know, we do shift our focus quite a bit. And I thought that this was almost like a good palate cleanser, so to speak, because we're talking about open source, and in particular, some big issues that have come up in open source communities and open source discussions. Recently, you know, I'm not going to claim to be some big expert on open source communities. Tante is much more involved in that than I am, but there has been quite a bit of controversy lately

around what has been happening with WordPress and a fight in particular that one of WordPress's founders, Matt Mullenweg, has been engaging in with a company called WP Engine, but also these broader discussions around AI and what open source AI actually looks like, what it means, and if it's really any better than, you know, any other kind of generative AI model. Because some of the assertions that we've been receiving are that closed AI is bad, but open AI is inherently good. And that if AI is open source, then there are not these big concerns that we have with so many of the other

generative AI models or just generative AI in general. But I'm not sure that that's the case. And so what I wanted to discuss with Tante in this episode was to dig into these key issues, these stories, you know, if you haven't been following them as well as we have, but also to talk about what this means for our understanding of open source and the role that this concept in this community and this approach to software actually means in 2024 as commercial pressures have not just been shaping the web and what we do online and seemingly eroding so many of the services that we rely on, but also are shaping the way that the open source community works and that we understand open source

in ways that are not necessarily so great. And whether that means we need to adjust our understanding and concept of open source, whether that means we need to approach it in a different way, or whether we need to think about a different way of framing a better approach to software that maybe is not so focused on licenses and understanding source code, but something else completely. So hopefully you find this episode interesting. Hopefully if you're someone who doesn't engage so much with open source communities or are not as informed on what that means that you learn from this episode, because I have to say I had a really great time talking to Tante about all of it. And he definitely brings his political perspective and his Luddite perspective

to our understanding of open source and what is happening with it right now. So if you do enjoy this episode, make sure to leave a five-star review on your podcast platform of choice. You can share the show on social media or with any friends or colleagues who you think would learn from it. And if you do want to support the work that goes into making the show every single week, so we can do things like make the Data Vampire series and have so many of these in-depth discussions on important issues that affect Data Vampire series and have so many of these in-depth discussions on important issues that affect your life and the lives of so many other people around us.

You can join supporters like Nelushi from Montreal, Dylan in Grand Rapids, Michigan, and Paul Louis in Amsterdam by going to patreon.com slash techwon'tsaveus, where you can become a supporter as well. Thanks so much and enjoy this week's conversation. Tante, welcome back to Tech Won't Save Us. Thank you for giving me a second opportunity to be on this wonderful podcast. Absolutely. You know, I always love to chat with you.

Even when we were talking earlier this year, I was like, I don't know how it's taken this long to get you on the show already. So having you back again so soon is like totally okay because you should have been on way earlier anyway. Yeah, I'm like on my long path following Brian Merchant to like become a regular guest here. Like, don't we all just want to be Brian Merchant? I mean, it's like the dream.

Yeah, we all want his long hair. Yeah, well, you're already on that. Yeah, but it's thinning, so. Yeah, well, I understand that. Yeah, yeah. But yeah, very excited to have you back on the show. And I think a really interesting topic, right? And something that we haven't really dug into, but I've been sort of following,

but probably not nearly as closely as you have. And that is like, you know, what is going on in open source. But I think to start like how that is represented by what we see going on with WordPress right now. And I think as many people know, WordPress is like this essential kind of publishing tool that so many websites rely on i think it's like something like 43 of websites use wordpress at least over 40 of the biggest 10 million websites i think okay gotcha but it's like very relevant like things many people might know ars technica they run wordpress like big big big publications run WordPress and small dinky blogs like mine or whatever. But it's, it's a lot. Yeah, I used to use it a lot. Like back in the day, I haven't used it, I would say in the past few years, but I used to build websites on WordPress,

like all the time up to about, you know, five or seven years ago when I just stopped building as many websites as I used to, because I used to get an idea and I would buy a domain name and make a website and probably never do anything with it after. But, you know, that was a fun thing to do for me for a while. So I used to be very familiar with WordPress. I haven't used it in a while, though. So when the story came up, I was like, wait, what is going on at WordPress? So maybe you can describe a bit more for us what is actually going on here. I believe this started in September. Tell us what's up. So like maybe just to name the main actors on this stage. WordPress is an open source project.

As many other open source project people might know, you can download the software for free. You get the source code and you can install it on your very, very cheap web space. And you have a publishing thingy going, which is one of the reasons why WordPress was so popular because it runs on every cheap, shitty little host. And it has plugins for everything. Whatever you need, there is a plugin for it. Some you might have to pay for, but the whole infrastructure you get by using the WordPress software is amazing. It's crazy what you can get there. And for someone who doesn't know web design, there's also like so many themes out there. So you can make your website look so many different ways without having to know how to

like go in and change the code and all that too. Right. Exactly. And you don't have to spend like, yeah, you probably have to spend a few bucks a month on your hosting and the domain or whatever, but it's very, very cheap. And still like, I mean, every software sucks. Every software that is used sucks, but it's decent. It does a great job for like you want to type a thing and you want it to appear on your website. That's great. So WordPress, many people believed, was owned by the WordPress Foundation, like a nonprofit that people thought owns like the trademark and the wordpress.org domain where your own WordPress installation

asks for plugin updates. This is where you can find all the plugins and all this stuff. People thought, okay, the WordPress foundation is what governs this infrastructure for the common good, if you want to call it that. Like this open source project and like they provide the infrastructure

for the source code hosting. They also sometimes help people who work on on wordpress with like travel budgets and all kinds of stuff like what these foundations for open source projects do then there is automatic automatic is a company owned by matt mullowick matt mullowick is one of the founders of wordpress of the wordpress open source project which again like we don't want to get into the code so much, but WordPress itself was a fork of an older platform that existed before. That wasn't as successful,

but like WordPress kind of took off for a bunch of different reasons. But Matt Mullenweg founded like this WordPress.org foundation that was the open source arm. And then his company, Automatic, that runs WordPress.com, where you can pay to have your wordpress hosted for you like you don't even have to rent web space and domain whatever you just go to wordpress.com say hey i want this domain registered for me run the thing do the thing i just want a window where i can type shit and it's supposed to appear which is a very convenient offer like people like that

people don't want to dig around with uploading php files to an ftp server anymore which totally makes sense like valid business model made matt molyneux really really rich like he's well off like he doesn't need to give a shit about anything anymore like not mark zuckerberg level rich but like he's got a lot but still still very very comfortable very very comfortable like he's he's the the guy that when tumblr basically crumbled he bought the rest of tumblr and keeps that running for whatever and that all belongs to automatic all kinds of other like surrounding platforms also belong to automating some people might know gravatar which gets your icon your avatar appearing everywhere that's also

run by automatic i was on there for the first time time in years the other day because I was trying to figure out how to make my avatar in ghost comments appear. And I was like, where do I put it? So I looked it up and I was like, it's Gravatar. I haven't used that in so long, but yeah. Yeah. Like Automatic runs a lot of that infrastructure and a lot of that for free. You can use Gravatar for free. You don't need to pay. You can also get a free free blog on wordpress.com you don't have to pay like there's probably some ads in there and you don't have all the features whatever but like the usual spiel as you know it but there are competitors that also offer you hosting your wordpress for you and also like a managed hosted wordpress like automatic does where you don't even have to care about having a cert whatever like just you yell at a person and say give me a WordPress. And they do that. And like one of those competitors of Automatic is called

WP Engine. They also offer you to run your WordPress for you. They are probably about as big as Automatic when it comes to how many blocks they host and whatever, but like they do the same thing. And there's a whole bunch of other companies who do that for you. They might not be as big, but there's a lot of them. Basically every web host that you go to today has a button for like, just give me a WordPress. And they just give it to you for a few bucks a month because WordPress is so established and so relevant. Like people want just that. Like for some people, that's just how you publish on the internet. So like those are the players that we have. For a long time, everyone was kind of fine it felt like everyone was fine fine like wordpress was chugging along like updates every

few months new features whatever blah blah blah automatic made a lot of money wp engine made a lot of money but for some reason and we don't know why in september matt mullenwick was no longer okay with this. So it's sometimes a bit hard to fully reconstruct what exactly happened when, because you get a lot of the information just through leaks and whatever. But I'm trying to be as neutral as possible here. Matt Mullenweg basically sent over WP Engine a demand for money. The argument was they weren't contributing enough to WordPress, which I think might be true. You could probably argue that many companies who use open source software

to make a lot of money are not really contributing a lot. Knock on Jeff Bezos' door and ask how much Amazon is really contributing to all the open source software they are running for everyone and making a lot of money with. So he might have a case here. We could always argue WP Engine should maybe do more, hire more developers or whatever. Like, okay, this is a conversation that you can have. And just to be clear, when we're saying not contributing enough to WordPress, we're talking about WordPress, the open source project,

WordPress.org, not WordPress.com, which is the commercial entity run by Automatic, where you can host your website. Yes. Like his surface level argument was, I'm putting all these hours of Automatic employees into the open source project to make this better that you profit off and you are not giving back. Like you're not contributing as much as I am, which again, might be true. WP Engine is owned by a private equity firm. They are not in it for the good of everyone. They want to get out as much money as possible.

Okay, that's how companies work today. That's the game. You're saying private equity isn't in there for the public good? Private equity is also in automatic, but we're not talking about that. Yeah, let's not talk about that piece. Exactly. But on the other hand wp engine did support the wordpress community like they funded

a lot of conferences and like they were like the top sponsor in many many wordpress conferences to like make them possible so it's not like they only took so i guess their argument would be like okay we don't contribute as many hours to the open source project that is WordPress.org and, you know, the WordPress project, you know, creating this software and creating this kind of ecosystem of services. But we contribute to the WordPress community in other ways by funding these conferences, funding these initiatives to try to get people more involved in WordPress and all this kind of stuff. I guess that would be kind of what they're saying. They're also funding like actual development. Like we talked about like WordPress has all these plugins they're called, like new features for the thing.

And a few really core plugins were developed by WP Engine or the developers were paid by WP Engine to get this thing going. So it's not like they didn't do anything. But again, I'm totally on board arguing that they probably could have given more. Like they made a lot of money, so they should share, but they're not. Okay, whatever. But what Matt Mullenweg demanded was not just, hey, please contribute more. Basically, he asked, I want 8% of your gains paid to Automatic, which looks weird because

Automatic isn't the open source project. Yeah. Sounds strange. Sounds strange. Okay. I mean, everyone can demand anything. I can also send Matt Mullenweg an email demanding 8% of his gains for whatever, but like, okay, why should anyone care? Well, Matt Mullenweg speaks at big WordPress conferences. So he was basically threatening WP Engine with like scorched earth tactics. Say, okay, you're going to do this, or I'm going to shit talk you on every stage I can get to. And that's what he did.

He called them a cancer to the WordPress ecosystem. Including the stages that WP Engine is funding for him to be on. Exactly. Exactly. And he was at that conference where he did it, like a big WordCamp. He was actually seen basically attacking the WP Engine employees that had a booth there because they were a gold sponsor. And they don't know shit.

Like they run the booth there for whatever. So he talked shit about them, like getting out some weird excuses about them butchering the WordPress code, whatever. Like it didn't really make sense. It felt weird because the things didn't add up. Either your problem is they're not contributing to the open source project. Okay. Valid argument. Why are you demanding they give money to your company? That sounds strange. Okay. Yeah. And I guess this is where things start to pivot, right? Because you have this initial argument that this is about not contributing to the open source project.

But then as it evolves, it becomes, oh, you're using the WordPress branding or the WordPress name. And actually, this is a licensing problem, which Matt Mullenweg is using as the justification for demanding 8% of revenue and stuff, right? Exactly. The WordPress Foundation has a trademark on the term WordPress, and they have given out one commercial license to that, and that is to WordPress.com, to Matt Mullenbeck's company. They don't have a trademark for WP. That's not what they have. So the whole argument, again, is a bit fishy. And there's so many companies that have WP, whatever as the name, and that make money

off of the WordPress, of their wordpress services whatever it is like building plugins building themes there's so much going on but he picked them for some reason probably because they're the biggest company like also fun fact automatic used to be invested in wp engine wild they sold their stake to the private entity it's a confusing mess another detail i found really interesting there, because I was reading about this in The Verge as we were preparing for this, was on September 19th, the policy basically said that companies were free to use the WP abbreviation any way they saw fit. But that language was deleted after September 19th and saying that other companies couldn't

use WP in a way that was confusing to people and specifically called out WP Engine there. So it even seems like before Matt Mullenweg launched this whole campaign, what they were doing was perfectly fine. You know, we can debate as to whether they were doing enough to support the open source project, which I think is a legitimate thing, as you're saying. But then when it comes into this whole debate around the licensing and stuff, which is the justification that seems to come later from Matt Mullenweg, that is something that is created just ahead of this whole outburst that he has, so that he has some justification for

trying to demand this money, launching lawsuits, all this kind of stuff. Exactly. And the language changed that you just pointed at, that was on the wordpress.org page. That was on the foundation site. That wasn't Matt Mullenweg's personal website or whatever we believed, but it was on the organizational. That should be separate things. And to be clear, it also says in that policy that if you would like to use the WordPress trademark commercially, please contact Automatic, which is the commercial company, not the organization. Exactly. It's a strange mess that just certain people seem to do things and roles that they shouldn't have or that they shouldn't act in this way. The WordPress Foundation is this abstract thing that runs the project, whatever that means. Who runs the foundation? It has three directors.

One of the directors is Matt Mullenweg two other people there basically have nothing to do with the community like they are not developers they they haven't a long stake in the community like one guy did wordpress tutorials till like 2014 or whatever when his company was bought by matt mullenweg's company and now he's on the board for whatever reason so it became painfully obvious that the word WordPress foundation that people believed was in charge of this open source project that basically everyone relies on, like either directly or like because the people that publish the things that you want to read use it, that this foundation really isn't a foundation, but that's just one guy who also has a company that tries to make money off the thing that the foundation does and who seems to switch roles whatever way he sees fit at any given point in time. And that kept on happening. A bit later, when the conflict escalated a bit, lawyers were sent out and cease and desist and all that kind of stuff.

At some point, Matt Mullenwick, probably as the director of WordPress.org, decided that WP Engine's customers no longer could access the WordPress.org infrastructure, which means no one who ran their blog or their website on WP Engine could get security updates for their software anymore. No one could install new software anymore because that always goes directly through WordPress.org because that's where you're supposed to get the canonical right and updated versions of everything. That's like WordPress.org is the trust anchor of the whole infrastructure. That's where everything lives because the foundation runs it.

That's why people believe that this is a trustworthy enterprise. Everyone can get you some code to install on your website, but you never know if that makes sense or if that's trustworthy. But if it's on WordPress.org, you have a certain level of trust to it. He basically disconnected everyone on his competitor's website from that infrastructure where there was never a terminology that WordPress.org wasn't for everyone who wants to use it. There isn't a way to like, I need to pay for it. You don't even need to have a login there basically to install stuff.

That's the reason why WordPress got so big, because you had this big repository, themes, plugins, whatever, that you could just install. But suddenly for some people that no longer was true, which was kind of a break with the social contract of open source. What's going on here? What happened? He even escalated that. The people who have WordPress.org accounts, because they are developers of the open source software or they're contributors to the project, they run the accessibility team or the social media or they run events, whatever, they have a WordPress.org account, on the beginning of October, a weird little checkbox appeared under the login form where you enter your email, enter your password,

and then there's a checkbox where you have to say, I'm not in any way affiliated with WP Engine. So you can only log into the system if you basically legally claim not to have any affiliation with WP Engine. What does that mean? Like if you're a freelancer and you have clients that run their stuff on WP Engine, are you no associated with them, affiliated? This wasn't something that the open source project decided.

There was something Matt Mullenweck decided. And everyone was surprised. Like the people in the open source project were surprised. Like in an internal slack of the project, they asked him, okay, what does this mean? If I just check it and it's not true, will I get sued? And he just refused to answer basically. It's like, yeah, you got to know what you're doing. Okay.

Thank you. I mean, it was so bad that the people who run like the accessibility team that tries to develop accessibility features for, they couldn't meet because no one who manages the team could log in. Like that's the state where he left this project going. The checkbox is still there. That's so wild. So I guess based on that, I kind of have two questions, right? You know, you're talking about this, this big disruption at, as we've been saying, this, you know, kind of piece of software that is so foundational to so much of the web, or at least the web that most people use, right? Because so many websites are built on top of it. So I wonder, first of all, what is the fallout of this for WordPress itself?

Is this project, as we understand it, kind of threatened? What does this mean for the future of it? And then I guess the broader question that comes out of that too, after addressing WordPress specifically, is what does this tell us about open source more broadly and open source projects more generally? I mean, it's hard to say what this will do to the project because many people who use WordPress don't really know a lot. Like they just hired a dude or a gal to build a website for them. And they use WordPress because many people do, but they're not really involved with the whole thing. They just know, okay, I didn't have to pay a license for the thing.

That's cool. But for the WordPress project, what we see is that many people who had influential roles, who were contributors for many, many years, who developed core features or who maintained core plugins that basically everyone used publicly left the project saying okay i'm not no longer contributing to this matt mullenweg because he also got criticized by his employees he gave everyone a buyout offer like okay at least 30 000 bucks just tell me you don't agree with me fine you can go here's your money keep your laptop i don't give a shit he even upped the ante like a few weeks later saying, hey, here's nine months of a buyout

and you can just leave. He lost a lot of people. We know as of early October when that first offer went out, it was 159 employees or 8.4% of Automattic's staff that took it and left. And there's been recent reporting that suggests the number is quite a bit higher now. Exactly. And of course, some of them will be working on WordPress.com.

So, I mean, that's his business risk. If he wants to treat his employees badly, that's his thing. But of course, this will also decrease the amount of resources that the WordPress project will have at their disposal,

which is a challenge. But also like WordPress also lived off, as I said, like people use WordPress who don't know, who didn't choose that, who just use it because somebody who set it up for them use it. What's with these people who made a living setting that up for people? Can they trust this platform anymore?

They could be locked out of their source of income tomorrow because they criticized Matt Mullenweg on Twitter. That can happen. That's a thing that can happen. Well, you can only imagine so many people who, as you say, are not technical people who wanted to get a website set up and hired someone to help them do that. And maybe the person who they hired to help them do that decided to put it on WP Engine. And now all of a sudden, their website is threatened and they don't know what to do about it or why there's this problem

in the first place. You know, it was just supposed to be this dependable as you say trustworthy infrastructure and now all of a sudden that is kind of thrown into jeopardy and even if you're not on wp engine the question then becomes i guess like what is the future of this project if this guy matt mullenweg is just kind of like losing it and has total control over it yeah and that is i think very dangerous for that community like wordpress isn't where it is because it's the most amazing piece of software ever written. Yeah, I've heard it's really bloated now. Yeah, it's very old.

And I'm not saying it's bad, but like code that old just gets a few weird idiosyncrasies and it gets Byzantine because you want to support the old versions because people have surrounded, whatever. It's hard to run a project that long so the code quality or whatever was never the point for wordpress the point for wordpress was there's this trusted in this trusted infrastructure that i can hook into where i can get all kinds of stuff that is open source that i can run for myself like that is something that we've not like the two of us,

but like we as like the tech people, that's what you've been telling people. Like you need to get your own domain, run your own website, publish to your own control, do your own thing. And people did that. And they used WordPress

because that's one of the most affordable ways to do it. There are many other more fancy technologies people use these days, but they are significantly more complex to run and oftentimes more costly. And you need to know more to run it, or you have to pay a company to do it for you. And that is a big challenge for the role that this project has, because what this creates is of course, friction. And I mean, some people leave the project, but they still kind of want to support their customers. So there's always like this, this conversation with an open source

called a fork. Because the code is open source, you can take it and develop it on your own in a different path or with a different governing body. Like that happens a lot. Like you said, with what WordPress did originally, right? It seemed like it forked something

that was already there. It was not an unfriendly fork. Like the developer of the original software was also part of WordPress. Forks happen. But if you fork now WordPress into like, there already is an older WordPress fork for some reason, but that never really took off. But now you have like three, four, five different versions. And in like five years, you suddenly no longer have like this large platform that everyone can use. But you're in some splinter that certain

features don't work. And you don't even know anymore because you just thought, I installed this WordPress thing and it always worked and it no longer does. So for the project interpretation, this is a real danger. As well for the professionals who live off of doing WordPress, but also for people who kind of rely on it and who never know what happens next. And of course, if you have like, say you have users or you run your own web shop for whatever, which you can do with WordPress, which works. Now you no longer get security updates for some reason. Or Matt Mullenweg decides, okay, if you make money with this, I'm going to try to force you to pay me a cut. No one knows what's going on. Mullenweg isn't predictable in what he's doing. Like it didn't make sense to go this whole scorched earth path,

but that's where we are right now. And what came out in the conversation, no, WordPress.org doesn't belong to the foundation. That's his private website. He owns the domain. That's his thing. And he actually argued that he didn't have any responsibility to host plugins that WP Engine developed and paid for on the WordPress.org website because he said, hey, that's my website. I can decide whatever I publish there. Yeah, that seemed really notable as well, right? That as part of all this, he basically said, WordPress is mine, not just the commercial side of it, but, you know, the kind that's supposed

to be an open source project too. And I wonder, you know, pivoting then, like getting to the second part of my question there, what does this mean for the broader discussions that are happening now about open source and the future of open source to see such a notable and such a large and significant open source project, you know, becoming as unstable as it has become, as you've been saying? Yeah, the license of the code hasn't changed. WordPress is still open source. It's a GPL license project license project you can download it you can do whatever you want with the code that's the rules but what this has shown us again like for many that might not be a super new uh thing but like this makes it so tangible how little this whole licensing thing does for us like it's not nothing like open source licenses are important they're an important tool but they only get us so far because many of the even big open source licenses are important they're an important tool but they only get us so far

because many of the even big open source projects they have this there's even a meme that some projects are run by like bdfl a benevolent dictator for life and that's not said as a bad thing that's like oh it's cutesy linus torvalds is a benevolent dictator for life for linux because linux is what linus torvalds says linuxent dictator for life for Linux, because Linux is what Linus Torvalds says Linux is. And many other projects like the Python language, which is like the language I mostly use, they had that for a long time until that guy said, this isn't cool. We need to find a different strategy where people have like stake and there's voting going on and transparency and it's not up to my whims.

But many projects still have like this very strong man. I know of no project where it's a woman that takes that role. It might exist. I know of none. So you have these strong men that, as long as they are fine, as long as they are mentally stable, everything works. But then, for some reason, things shift, and you never know why. Suddenly, things can spiral very quickly. I think that's what we've seen with WordPress. Like people didn't, you could have seen that something was up with the way that this project was organized. A project that, as we said,

runs more than 40% of the web, basically. And a lot of that is spam as well, but okay, like it runs so much and no one ever asked, okay, why is the foundation run by the guy who also has this company? Why is there no run by the guy who also has this company? Why is there no election of the directors?

How is the community really having a say in what's going on? Like, because Automatic pays many of the developers, Automatic also decides what gets developed. Which, I mean, that's how open source works. Who writes the code decides what goes. But many other projects try to, like, structure, okay, what is our roadmap? What are our goals? What are are our values what do we want and what would be for example something that that automatic could develop as a plugin for their extra customers why does this need to be in the core wordpress package like the core wordpress package if you just download the thing from the wordpress.org it does a lot of upselling for automatics products like automatic has like a jetpack plugin that does all kinds of shit for you and they try to

weasel that in yeah that's what immediately came to mind when you were saying that yeah totally um the automatic has this very special relationship which i mean i understand like the founder and the company like but if you are a professional who wants to govern the infrastructure that runs so much of the web like at some point you should ask why is this happening this looks weird as you were describing the benevolent dictator part of it i was thinking of like in some countries that like have dictators when the dictator dies all like the fighting over who's going to have power next like that was what was going on in my mind like who is going to take over the throne of the open source project when the benevolent dictator founder of the project dies yeah like

many projects have evolved and have established like actual grown-up processes where there's like people are elected and there's term limits and transparency like democratic because an open source project of a certain size needs to be democratic to make sense. So the users, the developers, they all need to have like the ability to have a stake in this. But WordPress never had that.

But as long as like automatic kept paying some of the bills, no one really looked at it. Like Matt is such a nice guy. It's fine. And for a long time, like many people saw him like like that like he gave money to all kinds of courses like he bought tumblr when it would have been shut down otherwise but like even a few months ago like he started selling all the content from wordpress.com and from tumblr to ai companies to train their models on to make some money like that was the first rift where people realized maybe he's not such a cool guy

but like i mean he has a company he wants to make money it's fine and he pays a lot of people which is also good like people need to make a living i get that maybe wordpress just needs to become a dao a decentralized autonomous organization and bring us into the web3 future like that seems that that is that is one path or you could just look at what other similar projects have done but of course he would have to seize control of a lot of the stuff and things would maybe change maybe the project would develop in ways that wouldn't perfectly align with automatic strategy and he's not willing to give up that the recent interviews he's kept repeating like no no i'm not gonna gonna seize any sort of power over this project. This is my project, basically.

You're just guests. And that means the whole idea of trust into this open source project is broken. You can't trust this. You can't know what this guy is doing next. That is really something that should have been avoided a decade ago, maybe, or years ago, at least. And it shows how little even well-established open source projects think about this idea of participation and democratic principles that need to be in place for like these public infrastructures the idea of open source is always

collectively we're building this public good code for everyone to use for everyone to share like yeah the comments like we're contributing to the comments like public good code for everyone to use, for everyone to share. Like the comments, like we're contributing to the comments, like there's code for everyone and that's great. Okay. But maybe that's not enough to just throw code somewhere. Maybe a project is more than that. And maybe we need to think about governance and we need to think about how power works in this project.

Because if I control the software that you need to write, like your newsletter, your website, your whatever, I have a lot of power over you. If I can shut down remotely your website or can make your life harder, that is a lot of power and responsibility I have. And if that's not checked, that should make everyone very uncomfortable. Well, I feel like on that point, you know, it gets to something quite important, right? And discussions that have been happening for a while around the commercialization of open source, the way that it has been captured by a lot of these to be in the first place. Because as you were saying, it's so much about a license to the project, being able to see the source code to the project. But, you know, it's not the same thing as like free software, anyone can just kind of do what they want sort of a thing. How do you see that kind of broader trajectory and discussions and understanding of open source?

The term free software is older. That was basically coined by Richard Stallman, who is a despicable man who should never be let near a computer, a microphone or anything. I was going to ask if that concept seems like tainted by its association with him. I think it is. He's a horrible human being. But he came up with this idea. And I think for somewhat weird reasons, and I think they don't really hold up for the world today and for what I'm trying to argue for it. Okay. The idea of free software was that people who run code on

their machine or who have to interact with code should have the ability to change it, to see it, to change it, and to shape it in the way that it serves them, which is a liberating idea in general. And that was what he built the GNU project, it is called, and the Free Software Foundation around this idea. Like, okay, if you don't have the code to the software you run, you are controlled by the software it's a very simplified idea but for him it goes so far that he doesn't use laptops where the firmware like the stuff that programs the chips in there if that isn't free he's not going to use it like for the longest time he didn't use wireless network cards because they had proprietary firmware in there he goes that far like he's very extreme in that regard extreme Extreme to a fault,

probably. But the idea makes sense to say, okay, if our world is so built on us living at these parts of our lives and digital infrastructures, we should see what's going on there. We should know what's going on there. We should be able to change things in the system that we use to serve us. Because the computer that we bought should serve us because the computer that we bought should serve us it should help us do what we want and not block us from i don't know you've downloaded the software and i don't want to let you start it or whatever that's something that goes against what the free software movement went for but the free software movement even though its politics are a bit shady and i don't think we can get into it, it's very libertarian in its leanings. It was definitely,

at least a politically influenced movement. It was more than just a license. Business didn't like that, but business liked getting free code. Free labor is cool, but the whole politics are a bit shit. So the open source initiative formed to kind of like the idea of open source like this terminology was there before but like they published the open source definition what makes a thing open source that was based on debian's free software guidelines debian is a linux distribution also a very free software leaning distribution that if you don't change things in the settings it doesn't install things that are binary. You can't install Slack on there without much fiddling around

because they think, hey, this isn't what you want. You want something that is open for you. And the OSI, the Open Source Initiative, kind of codified what open source means as the open source definition, which you can find on opensource.org. They also got the domain, and they kind of bless licenses that you can put your software under as either open source or not and if it's open source like they are compatible you know if you have an mit license thing and an apache license thing you know okay i can just mash these things together and they have

no license trouble which corporations really liked like that's cool you know which software you can use for free. That's great. You don't have to pay the people who build it. It's not a very complex document, but it's very much based around the idea of licensing. Yeah, there's some political wording around it, but the main case is, is this license compatible with this? And what do you have to do if you use this license, whatever? That's what kind of shaped what open source means. But at some point, people no longer said free software. They started saying open source.

And that kind of was an important shift, I think, because it kind of allowed the whole conversation to forgot that there's politics behind this. Politics that should be analyzed and maybe updated a bit and changed and whatever. But there were politics and they've been not fully stripped, but at least taken a lot back. And I think that that was where some of the problems started. Because yeah, WordPress is an open source project. Yes, you can get the source. But we realize in this case, just getting the source isn't really it.

That's not enough. I guess that shift kind of happens like in the 90s. Would that be right to think about? I think it started already in the 90s. But I'm in Germany, as we know, I'm European. Germany is always a few years back. So it took a lot longer for things to reach here. So maybe in the US, it got relevant a bit faster. But these days, if your license is blessed by the OSI, even if legal documents talk about open source software, they reference what the OSI says open source is. This isn't a legally binding document, but it has quasi-legal power to a certain degree. Yeah, it's generally accepted. It's a very influential definition. It's massively influential. And it's not wrong. It's not that they made up

something different. It aligns with the mechanics that the free software movement came up with to support legally the things that they cared about, but they just took the tool and not the reasons why the tools are the way they are. And not fully, again, people there also, like the person who wrote the open source definition for Debian also contributed to the OSI, but there's people believing what they do and there's the way an organization influences the world by what it does. And I think the OSI, but like there's people believing what they do and there's the way an organization influences the world by what it does. And I think the OSI basically stripped all the inconvenient politics out of it. So you could still, and you could still say, Hey, this is open source. This is not open source. You could still be the good guy saying, Hey, we're contributing to open source.

You can have this thing from us, but it was a big shift. So based on what you're saying there, you know, in that discussion around the open source initiative, and its definition of what open source is, as I was saying, as and as we've been discussing, there has been this concern in the trajectory of open source and the ability of major tech companies to, you know, capture what is being done by open source to profit from it. And for the earlier kind of more idealistic hopes of what open source would be to not necessarily being followed through on. And I guess that plays into the other big open source story that we wanted to talk about. Because on the one hand, you have what is happening with WordPress and what Matt Mullenweg is doing.

But then on the other piece of this, there's been a lot of discussion over the past couple of years about open source AI, right? And this kind of assertion that if AI and generative AI in particular is open source, then it is like inherently better. And it's easier to address the problems with it, you know, as we've been discussing so much about open AI and the potential consequences of a lot of these large language models and things like that. And the open source initiative recently made a definition of what open source AI is. But there's been a lot of criticism of that definition. And it seems like, you know, kind of further degrading this earlier definition of open source and what it's supposed to mean. So can you fill us in on what's going on there? And what is so concerning about this development? Yeah, a few months ago, like,

maybe now it's even more than a year ago, the OSI started thinking about defining what open source AI is. Because, for example, Meta really pushed their models that they released as open source AI because you could download it and use it under certain conditions. Which I think is a really deceptive framing, personally. But yeah, it totally is. Like, we'll get to that in a second. And I mean, also, there's like open AI who kind of have that in the name but they're the opposite yeah

so i understand like their intention to feel like okay maybe we should clear this up a bit and then they started like a very collaborative process where many people who are kind of in the osi or in the around the osi like there's a public discussion forum where people could argue about the definitions and like there were many many versions and like on a few days ago like the actual 1.0 version like the the first official release was made in this their pitch was we have values that we have put into our open source definition. And now we want to apply these values to open source AI to define which open source AI is blessed by us and which not. And as you said, with this whole open source is always good kind of thing, because that's what everyone's been

hearing all the time. Like even civil rights oriented NGOs are saying, hey, everything should be open source like public money public code is a thing that's very big in germany where many ngos arguing hey if we pay it with tax money it should be open source and everyone should be able to see it why is this proprietary so the osi said okay we're gonna build an open source ai definition based on our values and like the the whole open source definition is based on four freedoms, they call it, that you should have as a user of a system. And the first freedom is to use the system in whatever way you want. You shouldn't have to ask for permission and the license can't tell you

you can use it for this, but you can't use it for this. You should be able to do whatever you want. Like an image generator model that would tell you you can't generate images of paris marx would not be open source it would go against what that said for the second freedom is you should be able to study how the system works that's where all the licensing thing comes from you need to be able to see the code to study how the system works that was why the whole licensing angle was so important. You should be allowed to modify the system to your needs. You should be able to take something out that annoys you or add something to it. And the fourth freedom is you should be

able to share that. If I take the code and I add something to it, I should be allowed to give that to everyone I want. Those are the four freedoms that the whole open source idea is built around. They said, okay, let's do that for open source AI. It worked very well for source code back then, back in the days, like when the whole open source definition was created. I was still a young student trying to get Linux to run on my laptops and wireless never worked. It was very annoying. But back then it really worked because you always got the right code. You downloaded packages that you could use and look at. And if it's proprietary, was very annoying but back then it really worked because you always got like code you download the packages that you could use and look at and if it's proprietary you remember like windows and you download like an exe file and ran it but you had no idea what it does whatever back then it

really worked this whole approach showed its problems already with the move to software as a service because all these software as a service products they use all kinds of open source but because they never give you source code like because you never interact with the system that way they can kind of cheat their way around giving you their modifications like they use the software and they obey by all the licenses but they kind of can cheat around they don't give you the full access. You can't go to Meta and tell them, hey, you use all kinds of open source in your stuff. Give me the code for the newsfeed or whatever, if anyone wants that.

I mean, it's just spam anyways. But we've seen there that it's a problem. And now they're trying to use that for AI. And AI is a bit of a different beast than source code. Source code is like you have a bunch of files, maybe a compiler runs it, and then you have a program. Like that's how these things work. With AI, you also have some code programs that run, but that's the smallest part.

A neural network is trivial. It's a few hundred lines of code. It doesn't really matter. But what matters is the data. Everyone knows what we call AI today is statistical model that has been trained on data, and it's like the extraction of patterns from that data. So you need to look at the data to know what's in there. The person who wrote the original open source definition, Bruce Perrins, also said for AI, the training data is like the code because you need it to understand it, to study the system, to see what biases are in there, to see why it reacts in a certain way. You need to be able to analyze the data.

But what the open source AI definition of the OSI did, like they said, okay, yeah, you have to give everyone the code. You have to give people the way the code train the data. And you have to give people the license to do everything with the network you give them. You have to give them the weights, which is called like, that's basically the statistical summary of the training data that you can use. You have to give people all that with all the freedom and they can share it and all the things that we talked about earlier. But they added that for the data, for the training data, you don't need to give people all the training data.

They said that you can also just describe the training data. You don't need to give people all the training data. They said that you can also just describe the training data. That's so funny. Just describe the data. You don't need to tell us what's in there. Just give us an idea. Yeah. I mean, they say you need to describe in a reasonable detail so people could reproduce something that is kind of like the network you have, but they can never reproduce the network you have. They never really know what's in there and what's not in there. They never fully understand the system. And one can say, this is just a purist argument because, hey, who will ever read all the data files and change that? But the point is, either you can understand what's in there and check it and filter it and see what's going on,

or you can't. That's been a bit of a challenge because they now basically labeled a lot of stuff that many people would consider proprietary open source. Yes, you can download like Meta's weights and run the thing, even though like the OSI wouldn't consider Meta stuff open source AI, because Meta restricts certain things that you can do. There's a few open source models, but you never know what's in there. It looks small, but it is a conceptual shift in whether you actually know what you're getting on your machine or not. And from a computer scientist point, you don't know if there's an attack vector in there. Many people who build actual software, they try to create reproducible builds. Like I give you the source code.

I show you how I built the binary. You can compare them and you can make sure that you get exactly what I told you is in your code. I can guarantee that you get exactly this thing running on your machine so you can fully trust whatever I gave you. This has been severed in this open source regard. And again, the OSI said, hey, we need to give some structure to the space. Open source AI is being thrown around and no one knows what it means.

So we gave them a definition. But this definition kind of gives corporations a lot of leeway to get away with things that no one beforehand would have accepted as open source. And the process leading there was the community told them that. Like for months, people told them, this is a problem. This isn't according to what we believe in as a community, as open source developers. But they went through with it. And they got a bunch of supporters on board, which is like mostly large language model producers who think this is a great definition of whatever it is. Mozilla is in there and a few others are in there who all think this is a great definition. I understand that because they want to be able to label their stuff as open source,

because that's basically like the software fair trade stamp that you can get. So we can say, yeah, we're one of the good guys, but no one argues about like, okay, but why can't they give you the data? And that is an interesting question. Like, why shouldn't you be able to give me your training data? And we know why that is, because a lot of that isn't licensed. Like a lot of that has been stolen. A lot of that you actually aren't allowed to use. A lot of that stuff has been maybe gathered in illegal ways and you shouldn't have it, but you can't show me that. You'll you'll just say yeah there's some video information in there because you don't want to tell them hey i scraped all of youtube

because then google's lawyers will come but we see again like the osi tries to build like a legal structure a legal looking structure like a structure of rules to give it a name and then it's open source but the community again is being like looked over and has this issue of okay what's the next step can i now release source code where i just give you part of the source code and the rest i just describe roughly yeah i wrote some really great code here and it really trust me bro it's not a crypto miner for source code no one would have accepted that but now that it's ai everyone accepts that and that's a big problem where even the the small values that we have in this open source community are undermined just because the tech changed a bit.

Right. That's what really stands out to me as I hear you describe that, right, is how the definition that is being constructed shows very clearly how the organization has been corrupted by the commercial impulses that we see corrupting so much else of what goes on online and what goes on in web communities and what is degrading so many of the services that we rely on. And even when you're talking about, you know, the description of the training data, it feels like in so many of these models, you know, we've been having this discussion for the past two years, in so many of these models that these companies are relying on the amount of data that they're using, and like the amount of computation that's required to train something like this is such that, okay, if you wanted to try to replicate it, like the number of people who would have the compute to actually be able to do that is so minuscule that, okay, even if they describe what the training data is, and even if you try to go gather it and try to go train your

own model and stuff, that's going to be so much more difficult because the power and because the amount of compute that you would need is not there. But there was one piece of this that I found really interesting that kind of relates to what you were just talking about, which is the legal element of this, right? You know, you're talking about how so much of this training data is stuff that has been scraped that you probably shouldn't have been able to scrape because it is, you know, under copyright or whatever. And you noted in a piece that you wrote about this, that in Germany, at least, and I'm sure a number of other countries, that if something is considered open source, you have more kind of leeway to scrape and to take things off the web than if it is considered

more of a commercial product. So if these models are increasingly considered open source, then that gives them a sort of kind of legal immunity or a bit of legal protection in scraping all the stuff from Disney or all the stuff from YouTube or whatever. I don't know about the other legislations, of course, but in Germany, there's like a scraping exemption for like copyright. If you do research, you're allowed to basically scrape everything, but you have to do research. But saying, hey, I released this as open source. This is my research project gives gives you access to these legal special rules that are supposed to be for universities and whatever. But suddenly you're able to use them. And then the big foundational model is then released there. And then your commercial thing is just twiddling around with the weights a bit, whatever, or

whatever. And I think this is a dangerous precedent of like a corporate takeover of a term that was supposed to codify like a common good like public good in a way like the commons in a software kind of way that is i think the most painful thing we've seen in the last few weeks and it's not a new thing like corporations have loved not having to pay people for their software work for a long time because like estimates think that reproducing the open source stack that we have today like commercially would cost in the trillions it's it's absurd how much money that would cost so yeah they love it and hey if if there's the comments then okay let them use it i don't i don't give a shit it's fine but the way that they start to wiggle their way into the comments and then try to like lock people out from it that is a really dangerous precedent i believe and i think that one of the reasons why

that worked so well is because of this shift from free software to open source where it has been like okay well we have this licenses and if they're compatible, everything is fine. And if a project wants to do some democracy and some more grown-up processes, fine, but others is also fine. If it's just a corporation and some guy named Matt, that's also fine. And I think we're seeing with these two examples that we haven't done our homework in the last decade because everything was going so well we had all the software and everything worked and every everyone gets all the software they need and now they can use i don't know whisper or whatever open ai releases as open source even though you just get weights it's it's just a just a binary blob it's a

proprietary tool that you can run for free. Like when I was young, you could run Winamp for free. You didn't have to pay for it, but you had no idea what was in there. And that's the same for these models. And I think that if as a society, we think that software should be in the core of so many of the services that we use, and we think that software can be part of maybe a better future or a more liberating future, can contribute to that, then I think we need to defend what that means. And I think that just looking at licenses and technicalities of which code can be linked to which code isn't enough. Saying this is an open source project doesn't cut it anymore. Maybe it

needs to be like a public good project, which it is open source but it also has democratic participation of the developers and the users and all the different stakeholders whatever like thinking more about these processes of governance of thinking okay how do we decide what happens which of course is hard because people are not paid for this a lot of the time people just do it in their free time and they work on what they care about and what they scratch their itches, what programmers say. So it's hard to tell them, hey, you need to work on this

because that's our strategy. That obviously doesn't work, but okay, how does such a community define their values and where they want to go and how, if not enforce, how can you get that done? If those are your values, someone should work on it.

If no one works on it, that's not your values. And that's also important to realize and to communicate to the users and i think that we've not done a great job and many i'm not saying no projects have like many projects have even in the space around just the cms thing that wordpress counts there's many projects that have done better jobs that have like actual governance structures but what they often forget is like for example you can vote if you're a developer or a core developer like if you've put in enough hours you get to vote which is a way to organize it i understand but on the other hand is that enough like shouldn't users have a say in how things work if it's a user-facing thing if wordpress is used by 40 of the web and so many people use it why can only people who can write

php code or css for the design why can only they contribute why can only they say something and realize that they say something and i think the next step towards, let's say, the evolution of free software, open source, whatever you want to call it, needs to look at that. How can users actually have a say and have a voice and have a seat at the table? But looking further, okay, if this is a public good, how do we actually make it democratic? Because I think it can't be if it's not. Then by accident, it might be for a second, but like we can never rely on it. And if it's an infrastructure,

just as I think WordPress is an infrastructure, it shouldn't be in the hands of a person. It should be in the hands of the people. If it runs 40% of the fucking web, then we need to take it away and we need to find ways to make that more reasonable and more aligned with the values that at least people living in democratic countries share. And I think that's important. I think it also aligns with what I think free software

was supposed to mean. I'm not a libertarian, but I think that that is a reasonable way to think about how this can actually contribute to a public good. And maybe it will get slower because democracy isn't efficient, but it doesn't need to be. That's not what it's for. It's there to give everyone a voice and have everyone heard and give everyone their rights. And I think that that is the homework that we need to do in the next few weeks and months and years

to salvage this idea of software that is there for everyone. The comments that i think are massively important but that we see are being taken over by other interests that are not getting more people in and getting more people access and understanding but like locking things away and closing things again in a certain way and it's sad to see so many like i have the website for for the open source ai definition all open and like the organizations that slide through that support it it's like mozilla what the fuck are you doing this is shouldn't happen that's something that many people have seen in the last few weeks

that's why we've seen so much agitation about not really super meaningful changes in wordpress like if you're not on wp engine it doesn't affect you that much. But I think it shows structural weaknesses that we have in the structure that is supposed to hold up the web, which is a thing that where we spend, I don't know about you, but I spent a lot of my life there. It's where I keep a lot of my stuff. Yeah, I completely agree with you. And I appreciate that perspective as well, right? To understand what is actually going on in this space with open source as, you know, we're seeing these growing commercial pressures affecting so much of what's happening a lot of this nostalgia for a way that the web used to be is often from the perspective of coders and programmers and people who can make these changes to code and often doesn't think about the impact of users or people who are like non-technical users of the web and making sure that they also have a voice in what the future of the web is going to be and to make sure that

they're still able to participate in it without having the skills to code things on their own. And I feel like when you're talking about that democratic element and bringing users in as well, I think that's a really important piece there too, to make sure that everyone is included in this discussion, and that we can think about what this non-commercial web or non-commercial way of developing these things is going to look like. That's not just about the people who are contributing their coding time, but about everyone else who's using it too i think that's like a very important point that the visions of the web were structured by dudes like me like computer scientists and programmers like the free software idea was built by stahlman

for programmers like he wanted to be able to fix a shit and he can't do that but many others can't and i think that just doesn't just apply to like software in a narrow sense if we think about currently there's this still this wibbly wobbly situation of post twitter slash x like how is the social web reshaping uh with this platform kind of dying or changing in the way it has but there's's still the problem of, okay, you can go to BlueSky. Now the BlueSky devs decide how your platform looks. Or you can go to Mastodon. So the Mastodon team decides how your platform looks.

Where is your democratic right as a user of that platform to influence? Of course, you can vote on issues, but who understands what a GitHub issue of a project is and how to formulate that in a way that the devs will not just throw it out. I think that while this might sound for some to like, okay, software, I don't care about software. I think it actually affects how our lives work. And maybe that's something that we need to drag this discussion out of like the software part.

It's like, no, no, we all live in these digital spaces. And if it's a digital space, we should have rights, whether we can code or not. And I think that that is the shift that we need to these important issues and fill us in on what is happening with open source at this, what feels like a really pivotal moment for it. Thanks so much for taking the time. Thank you for the invitation. Tante is a sociotechnologist, writer, speaker, and Luddite. Tech Won't Save Us is made in partnership with The Nation magazine and is hosted by me, Paris Marks. Production is by Eric Wickham and transcripts are by Bridget Palou-Fry. Thank you.